WO2012087104A1 - Intelligent load handling in cloud infrastructure using trend analysis - Google Patents
Intelligent load handling in cloud infrastructure using trend analysis Download PDFInfo
- Publication number
- WO2012087104A1 WO2012087104A1 PCT/MY2011/000089 MY2011000089W WO2012087104A1 WO 2012087104 A1 WO2012087104 A1 WO 2012087104A1 MY 2011000089 W MY2011000089 W MY 2011000089W WO 2012087104 A1 WO2012087104 A1 WO 2012087104A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- component
- system metrics
- cloud infrastructure
- load handling
- trend
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5083—Techniques for rebalancing the load in a distributed system
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2209/00—Indexing scheme relating to G06F9/00
- G06F2209/50—Indexing scheme relating to G06F9/50
- G06F2209/5019—Workload prediction
Definitions
- the present invention relates generally to a system and method of load handling in cloud infrastructure that applies trend analysis auto scaling on system metrics historical information to predict increasing or decreasing trend in order to act in advance to automatically provision virtual machines for application workload distribution without any human intervention .
- a system for load handling in cloud infrastructure comprising; at least an auto scaling server (2) at least a monitoring agent (6); characterised in that the said auto scaling server (2) consisting of at least an analysing component (2A), a predicting component (2B) and a checking component (2C).
- a method of load handling in cloud infrastructure comprising steps of, gathering and storing system metrics historical information by monitoring agent (6); retrieving system metrics historical information from the said monitoring agent (6) by an analysing component (2A); performing regression analysis on the system metrics historical information by the said analysing component (2A); using the output obtained from the regression analysis to extrapolate system metrics by a forecasted duration as specified in the configuration data in a predicting component (2B); performing condition checks in the checking component (2C) based on the output obtained from the said predicting component (2B) by comparing the predicted trend with system metrics thresholds set up in the configuration data input component (10); launching at least an additional new virtual machine (8) through a provisioning server (4) when the predicted trend exhibits an increasing trend above the system metrics upper threshold (14A) and registering the said additional new virtual machine (8); repeating the above step until the system metrics drops below the system metrics upper threshold (14A); terminating at least an existing idle virtual machine (8) through the said provisioning server (4) when the predicted trend exhibits a decreasing trend below the system metrics lower threshold (
- FIG. 1 shows a trend analysis auto scaling system architecture in accordance with the preferred embodiment of the present invention.
- FIG. 2- A shows a graphic illustration of system metric against time showing a predicted trend exhibiting an increasing trend scenario.
- FIG. 2-B shows a graphic illustration of system metric against time showing a predicted trend exhibiting a decreasing trend scenario.
- FIG. 3 is a flowchart showing the process of provisioning and terminating virtual machines by means of trend analysis of the present invention.
- FIG. 4 shows another aspect of trend analysis auto scaling system architecture in accordance with another embodiment of the present invention.
- the said trend analysis auto scaling system comprises at least an auto scaring server (2), at least a provisioning server (4), at least a monitoring agent (6), at least two virtual machines (8), at least a configuration data input component (10) (that stores input such as system metrics breach conditions and launch configurations entered or set-up by the user) and at least a load manager (12).
- the load manager (12) may be dispensed with as illustrated in FIG. 4 described below.
- the auto-scaling server (2) further consists' of at least an analysing component (2A), at least a predicting component (2B) and at least a checking component. (2C).
- the monitoring agent (6) collects and stores all system metrics historical information in real time, which includes but not limited to the CPU (central processing unit), memory, network and disk usage.
- the said analysing component (2A) of the auto- scaling server (2) accesses the said monitoring agent (6) to retrieve the system metrics historical information that had been gathered and stored in the monitoring agent (6) and performs regression analysis on the said system metrics historical information.
- the predicting component (2B) then utilises the result obtained from the said regression analysis to extrapolate the systems metrics by a forecasted duration of time as specified in the configuration data entered or set up by the user and stored in the configuration data input component (10) to achieve trend prediction results.
- the checking component (2C) uses trend prediction results to perform condition checks by comparing the predicted trend such as predicted trend of network usage with system metrics thresholds set up by the users in the configuration data input component (10).
- virtual machines take time to be provisioned, which could be more than an hour for many instances and during this time the already overloaded virtual machines of conventional models, not having a trend estimation technique capable of predicting the requirement of additional virtual machines in advance and take steps to provision same, will still have to continue handling the excess load.
- the present invention is provided with a trend estimation technique allowing trend analysis on system metrics historical information to predict whether the system metrics exhibit an increasing or decreasing trend.
- Such trend analysis will enable prediction in advance on the requirement of additional virtual machines (8) and make the necessary provision of virtual machines (8) so that the existing virtual machines (8) can be relieved of overload before any overloading occurs.
- the existing, terminated as well as additional new virtual machines will be indicated by the same reference numeral (8).
- the present invention based on trend analysis on system metrics historical information will be able to predict in advance the reduction in workload and accordingly reduce the number of idle virtual machines (8) associated to the system as soon as reduced workload actually occurs to save resources and cost. Using this trend estimation technique, auto-scaling can react to a forecasted condition as soon as the limit of overloading or under-loading strikes.
- the provisioning of additional new virtual machines (8) or termination of existing idle virtual machines (8) described above is achieved in the checking component (2C) and are based on thresholds of virtual machine metrics or system metrics including but not limited to CPU, memory, network or disk usage that are set up by the user in the configuration data input component (10).
- FIG. 2-A there is shown a graphic illustration of system metric against time showing a predicted trend exhibiting an increasing trend scenario.
- the checking component (2C) of the auto scaling server (2) will be immediately activated to increase the number of virtual machines (8) to be hosted in physical resources (not shown) through the provisioning server (4) based on launch configuration entered by the user in the configuration data input component (10). In this manner there is reduced delay in provisioning virtual machines (8) as its requirement is predicted in advance.
- the additional new virtual machines (8) will be registered with the load manager (12), which handles the application workload distribution in the virtual machines (8).
- FIG. 2-B there is shown a graphic illustration of system metric against time showing a predicted trend exhibiting a decreasing trend scenario.
- system metrics lower threshold 14B
- actual measurements based on system metrics historical information were employed to plot the graph to create the actual trend.
- the last actual decreasing measurement (17B) which is taken immediately before the actual trend reaches the system metrics lower threshold (14B) all measurements are predicted.
- the checking component (2C) of the auto-scaling server (2) will be immediately activated to terminate at least one existing idle virtual machine (8) and decreases the number of idle virtual machines (8) through the provisioning server (4). In this manner there is reduced delay in terminating idle virtual machines (8) as its redundancy is predicted in advance.
- the terminated idle virtual machine (8) is deregistered from the load manager (12) to stop further workload distribution to the terminated virtual machine (8) thereby conserving resources. This process is repeated until the system metrics rise above the system metrics lower threshold (14B). In between the scaling process, there is a cool down period where no scaling is allowed. This is to prevent positive feedback effect.
- FIG. 3 shows a flowchart illustrating the process of provisioning additional new virtual machines (8) and terminating existing idle or excess virtual machines (8) by means of trend analysis of the present invention.
- the process begins when the user sets up a breach condition and launch configuration in the first step indicated by reference numeral (22). Then in the second step indicated by reference numeral (24) the analysing component
- the analyzing component (2A) retrieves system metrics historical information from the monitoring agent (6). Then the analyzing component (2A) performs regression analysis using the said retrieved information in the third step indicated by reference numeral (26). Using the results obtained from the regression analysis in the third step (26), the predicting component (2B) extrapolates the systems metrics by a forecasted duration of time as specified in the configuration data entered by the user in the configuration data input component (10) to perform trend prediction in the fourth step indicated by reference numeral (28). Based on the trend prediction results the checking component (2C) uses the said prediction results to perform condition checks by comparing the predicted trend with system metrics thresholds in the configuration data entered by users in the configuration data input component (10) as illustrated in the fifth step indicated by reference numeral (30).
- the checking component (2C) will automatically terminate at least one existing idle virtual machine (8) as shown in the tenth step indicated by reference numeral (40).
- the checking component (2C) then de-registers the existing idle virtual machine (8) from the load manager (12) to stop further workload distribution to the existing idle virtual machine (8) as shown in the eleventh step indicated by reference numeral (42). In this situation the systems metrics will be directed as indicated by arrow "A" and the ninth to eleventh step will be repeated until the system metrics rises above the system metrics lower threshold (14B).
- load manager (12) may be dispensed with. This is illustrated in FIG. 4, another embodiment of the present invention where the load manager (12) for handling application workload distribution is not employed.
Abstract
The present invention relates generally to a system and method of load handling in cloud infrastructure that applies trend analysis auto scaling on system metrics historical information to predict increasing or decreasing trend in order to act in advance to automatically provision virtual machines (8) for application workload distribution without any human intervention wherein the auto-scaling server (2) comprises an analysing component (2A), a predicting component (2B) and a checking component (2C).
Description
INTELLIGENT LOAD HANDLING IN CLOUD INFRASTRUCTURE USING
TREND ANALYSIS . TECHNICAL FIELD OF THE INVENTION
The present invention relates generally to a system and method of load handling in cloud infrastructure that applies trend analysis auto scaling on system metrics historical information to predict increasing or decreasing trend in order to act in advance to automatically provision virtual machines for application workload distribution without any human intervention . BACKGROUND OF THE INVENTION
Workload of various applications varies significantly. Inefficient handling of high workload could result in downtime for the applications. In cloud environment, the applications are run on virtual machines. Most people over-provision the applications by having more virtual machines than sufficiently required to cater for the worst-case scenario to handle unexpected encounters with high workload. However this unexpected high workload may only happen during a small fraction of time throughout the
operation. Hence, this leads to waste of resources as during the majority of time the virtual machines are idle or under-worked.
In cloud environment, auto scaling is used to automatically provision virtual machines for application workload distribution. Virtual machines take time to be provisioned, which could be more than an hour for many instances. During this time, the already overloaded virtual machines will still need to continue handling the excess load. Therefore there is an imbalance in application workload distribution.
It would hence be extremely advantageous if the above shortcoming is alleviated by having a. proactive approach to load handling in cloud environment that automatically provisions virtual machines in advance without any human intervention by applying trend analysis on system metrics historical information to predict whether the system metrics historical information exhibit an increasing or decreasing trend and by using this trend estimation technique perform auto scaling enabling the forecasting of conditions in order to act in advance before any overloading or under-loading of system metrics strikes.
SUMMARY OF THE INVENTION
Accordingly, it is the primary aim of the present invention to provide a system and method for load handling in cloud infrastructure which utilises trend analysis to perform auto scaling enabling the prediction or forecasting of conditions in order to act in advance.
It is yet another object of the present invention to provide a system and method for load handling in cloud infrastructure which is capable to automatically provisioning virtual machines without any human intervention to relieve the virtual machines of imminent excess load before any overloading of system metrics strikes.
It is yet a further object of the present invention to provide a system and method for load handling in cloud infrastructure which is capable to automatically terminating virtual machines without any human intervention before any imminent under-loading of system metrics strikes.
It is a further object of the present invention to provide a system and method for load handling in cloud infrastructure which is capable of efficiently handling high workloads thereby conserving resources and reduces wastage.
Yet a further object of the present invention is to provide a system and method for load handling in cloud infrastructure which is capable of reducing downtime for applications.
Other and further objects of the invention will become apparent with an understanding of the following detailed description of the invention or upon employment of the invention in practice.
According to an embodiment of the present invention there is provided,
A system for load handling in cloud infrastructure comprising; at least an auto scaling server (2) at least a monitoring agent (6); characterised in that the said auto scaling server (2) consisting of at least an analysing component (2A), a predicting component (2B) and a checking component (2C).
In another aspect there is provided,
A method of load handling in cloud infrastructure comprising steps of,
gathering and storing system metrics historical information by monitoring agent (6); retrieving system metrics historical information from the said monitoring agent (6) by an analysing component (2A); performing regression analysis on the system metrics historical information by the said analysing component (2A); using the output obtained from the regression analysis to extrapolate system metrics by a forecasted duration as specified in the configuration data in a predicting component (2B); performing condition checks in the checking component (2C) based on the output obtained from the said predicting component (2B) by comparing the predicted trend with system metrics thresholds set up in the configuration data input component (10); launching at least an additional new virtual machine (8) through a provisioning server (4) when the predicted trend exhibits an increasing trend above the system metrics upper threshold (14A) and registering the said additional new virtual machine (8);
repeating the above step until the system metrics drops below the system metrics upper threshold (14A); terminating at least an existing idle virtual machine (8) through the said provisioning server (4) when the predicted trend exhibits a decreasing trend below the system metrics lower threshold (14B) and deregistering the said existing idle virtual machine (8); repeating the above step until the system metrics rises above the system metrics lower threshold (14B).
4. BRIEF DESCRIPTION OF THE DRAWINGS Other aspect of the present invention and their advantages will be discerned after studying the Detailed Description in conjunction with the accompanying drawings in which:
FIG. 1 shows a trend analysis auto scaling system architecture in accordance with the preferred embodiment of the present invention. FIG. 2- A shows a graphic illustration of system metric against time showing a predicted trend exhibiting an increasing trend scenario.
FIG. 2-B shows a graphic illustration of system metric against time showing a predicted trend exhibiting a decreasing trend scenario.
FIG. 3 is a flowchart showing the process of provisioning and terminating virtual machines by means of trend analysis of the present invention.
FIG. 4 shows another aspect of trend analysis auto scaling system architecture in accordance with another embodiment of the present invention.
DETAILED DESCRIPTION OF THE DRAWINGS
Throughout this document, unless otherwise indicated to the contrary, the terms "comprising", "consisting of" and the like are to be construed as non-exhaustive, or in other words, as meaning "including, but not limited to".
In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be understood by those or ordinary skill in the art that the invention may be practised without these specific details. In other instances,
well known methods, procedures and/ or components have not been described in detail so as not to obscure the invention.
The invention will be more clearly understood from the following description of the embodiments thereof, given by way of example only with reference to the accompanying drawings which are not drawn to scale.
Referring to FIG. 1, there is shown a trend analysis auto scaling system architecture in accordance with the preferred embodiment of the present invention. The said trend analysis auto scaling system comprises at least an auto scaring server (2), at least a provisioning server (4), at least a monitoring agent (6), at least two virtual machines (8), at least a configuration data input component (10) (that stores input such as system metrics breach conditions and launch configurations entered or set-up by the user) and at least a load manager (12). Alternatively the load manager (12) may be dispensed with as illustrated in FIG. 4 described below. The auto-scaling server (2) further consists' of at least an analysing component (2A), at least a predicting component (2B) and at least a checking component. (2C). The monitoring agent (6) collects and stores all system metrics historical information in real time, which includes but not limited to the CPU (central processing unit), memory, network and disk usage. The said analysing component (2A) of the auto- scaling server (2) accesses the said monitoring agent (6) to retrieve the system metrics historical information that had been gathered and stored in
the monitoring agent (6) and performs regression analysis on the said system metrics historical information. The predicting component (2B) then utilises the result obtained from the said regression analysis to extrapolate the systems metrics by a forecasted duration of time as specified in the configuration data entered or set up by the user and stored in the configuration data input component (10) to achieve trend prediction results. The checking component (2C) then uses trend prediction results to perform condition checks by comparing the predicted trend such as predicted trend of network usage with system metrics thresholds set up by the users in the configuration data input component (10). As is known, virtual machines take time to be provisioned, which could be more than an hour for many instances and during this time the already overloaded virtual machines of conventional models, not having a trend estimation technique capable of predicting the requirement of additional virtual machines in advance and take steps to provision same, will still have to continue handling the excess load. Unlike such conventional models which are unable to predict as and when additional virtual machines (8) are required, the present invention is provided with a trend estimation technique allowing trend analysis on system metrics historical information to predict whether the system metrics exhibit an increasing or decreasing trend. Such trend analysis will enable prediction in advance on the requirement of additional virtual machines (8) and make the necessary provision of virtual machines (8) so that the existing
virtual machines (8) can be relieved of overload before any overloading occurs. The existing, terminated as well as additional new virtual machines will be indicated by the same reference numeral (8). Likewise the present invention based on trend analysis on system metrics historical information will be able to predict in advance the reduction in workload and accordingly reduce the number of idle virtual machines (8) associated to the system as soon as reduced workload actually occurs to save resources and cost. Using this trend estimation technique, auto-scaling can react to a forecasted condition as soon as the limit of overloading or under-loading strikes. The provisioning of additional new virtual machines (8) or termination of existing idle virtual machines (8) described above is achieved in the checking component (2C) and are based on thresholds of virtual machine metrics or system metrics including but not limited to CPU, memory, network or disk usage that are set up by the user in the configuration data input component (10). There are at least two system metrics thresholds (or breach conditions) being set up by the user in the configuration data input component (10); a system metrics upper threshold indicated by a bold line (14A) [Refer to FIG. 2-A] and a system metrics lower threshold indicated by a bold line (14B) [Refer to FIG. 2-B]. Referring now to FIG. 2-A, there is shown a graphic illustration of system metric against time showing a predicted trend exhibiting an increasing trend scenario. Prior to reaching the system metrics
upper threshold (14A) actual measurements based on system metrics historical information were employed to plot the graph to create the actual trend. After the last actual increasing measurement (17A) which is taken immediately before the actual trend surpasses the system metrics upper threshold (14A), all measurements are predicted. When the system metrics historical information exhibits an actual increasing trend as indicated by the increasing diagonal line (15 A) there will come a time when the system metrics upper threshold (14A) is breached after the increasing diagonal line (15A) surpasses the upper intersection point (18A) [the point where the increasing diagonal line (15 A) cuts the system metrics upper threshold
(14 A)]. The trend for a specified duration of time in future after the last actual increasing measurement (17A) is predicted and hence the increasing dotted diagonal line after the last actual measurement (17A) is referred to herein as the predicted increasing trend indicated by the reference numeral (16A). When such predicted increasing trend (16A) is forecasted in advance, the checking component (2C) of the auto scaling server (2) will be immediately activated to increase the number of virtual machines (8) to be hosted in physical resources (not shown) through the provisioning server (4) based on launch configuration entered by the user in the configuration data input component (10). In this manner there is reduced delay in provisioning virtual machines (8) as its requirement is predicted in advance. The additional new virtual machines (8) will be registered with the load manager (12), which
handles the application workload distribution in the virtual machines (8). This process is repeated until the system metrics drops below the system metrics upper threshold (14A). Referring now to FIG. 2-B, there is shown a graphic illustration of system metric against time showing a predicted trend exhibiting a decreasing trend scenario. Prior to reaching the system metrics lower threshold (14B) actual measurements based on system metrics historical information were employed to plot the graph to create the actual trend. After the last actual decreasing measurement (17B) which is taken immediately before the actual trend reaches the system metrics lower threshold (14B) all measurements are predicted. When system metrics historical information exhibits an actual decreasing trend as indicated by the decreasing diagonal line (15B) there will come a time the system metrics lower threshold (14B) is breached after the decreasing diagonal line (15B) surpasses the lower intersection point (18B) [the point where the decreasing diagonal line (15B) cuts the system metrics lower threshold (14B)]. The trend for a specified duration of time in future after the last actual decreasing measurement (17B) is predicted and hence the dotted decreasing diagonal line after the last actual decreasing measurement (17B) is herein referred to as the predicted decreasing trend indicated by the reference numeral (16B). When such predicted decreasing trend (16B) is forecasted in advance, the checking component (2C) of the auto-scaling server (2) will be immediately activated to terminate at least one existing idle virtual machine (8) and
decreases the number of idle virtual machines (8) through the provisioning server (4). In this manner there is reduced delay in terminating idle virtual machines (8) as its redundancy is predicted in advance. The terminated idle virtual machine (8) is deregistered from the load manager (12) to stop further workload distribution to the terminated virtual machine (8) thereby conserving resources. This process is repeated until the system metrics rise above the system metrics lower threshold (14B). In between the scaling process, there is a cool down period where no scaling is allowed. This is to prevent positive feedback effect. FIG. 3 shows a flowchart illustrating the process of provisioning additional new virtual machines (8) and terminating existing idle or excess virtual machines (8) by means of trend analysis of the present invention. The process begins when the user sets up a breach condition and launch configuration in the first step indicated by reference numeral (22). Then in the second step indicated by reference numeral (24) the analysing component
(2A) retrieves system metrics historical information from the monitoring agent (6). Then the analyzing component (2A) performs regression analysis using the said retrieved information in the third step indicated by reference numeral (26). Using the results obtained from the regression analysis in the third step (26), the predicting component (2B) extrapolates the systems metrics by a forecasted duration of time as specified in the configuration data
entered by the user in the configuration data input component (10) to perform trend prediction in the fourth step indicated by reference numeral (28). Based on the trend prediction results the checking component (2C) uses the said prediction results to perform condition checks by comparing the predicted trend with system metrics thresholds in the configuration data entered by users in the configuration data input component (10) as illustrated in the fifth step indicated by reference numeral (30).
In the event it is found that the system metrics upper threshold (14A) has been breached upon condition checking conducted in the checking component (2C) as shown in the sixth step indicated by reference numeral (32), that is the predicted trend exhibits an increasing trend scenario, new and additional virtual machines (8) will be automatically launched and added by the checking component (2C) as shown in the seventh step indicated by reference numeral (34). The checking component (2C) then registers the new additional virtual machine (8) with the load manager (12) to enable application workload distribution to be handled as shown in the eighth step indicated by reference numeral (36). The systems metrics will be directed as indicated by arrow "A" and the second to eighth step will be repeated until the system metrics drops below the system metrics upper threshold (14 A). In the event it is found that the system metrics upper threshold (14A) has not been breached upon condition checking conducted in the checking
component (2C) then it will check whether the system metrics lower threshold (14B) has been breached as shown in the ninth step indicated by reference numeral (38). In the event the system metrics lower threshold (14B) is found to have been breached that is the predicted trend exhibits a decreasing trend scenario, the checking component (2C) will automatically terminate at least one existing idle virtual machine (8) as shown in the tenth step indicated by reference numeral (40). The checking component (2C) then de-registers the existing idle virtual machine (8) from the load manager (12) to stop further workload distribution to the existing idle virtual machine (8) as shown in the eleventh step indicated by reference numeral (42). In this situation the systems metrics will be directed as indicated by arrow "A" and the ninth to eleventh step will be repeated until the system metrics rises above the system metrics lower threshold (14B).
In the event it is found that the system metrics upper as well as the lower thresholds (14A) (14B) have not been breached upon condition checking conducted in the checking component (2C) then the systems metrics will be directed as indicated by arrow "B" and the second to eighth step will be repeated.
Alternatively the load manager (12) may be dispensed with. This is illustrated in FIG. 4, another embodiment of the present invention where the
load manager (12) for handling application workload distribution is not employed.
While the preferred embodiment of the present invention and its advantages has been disclosed in the above Detailed Description, the invention is not limited thereto but only by the spirit and scope of the appended claim.
Claims
1. A system for load handling in cloud infrastructure comprising; at least an auto-scaling server (2) at least a monitoring agent (6); characterised in that the said auto-scaling server (2) comprising at least an analysing component (2A), a predicting component (2B) and a checking component (2C).
2. A system for load handling in cloud infrastructure as in Claim 1 further characterised in that at least a load manager (12) is provided to handle application workload distribution to virtual machine (8).
3. A system for load handling in cloud infrastructure as in Claim 1 or 2 further characterised in that a configuration data input component (10) is provided to receive pre-determined inputs such as breach conditions and launch configurations from a user.
4. A system for load handling in cloud infrastructure as in Claim 1 or 2 wherein the said analysing component (2A) retrieves system metrics historical information from the said monitoring agent (6) and performs regression analysis.
5. A system for load handling in cloud infrastructure as in Claim 4 wherein the said predicting component (2B) extrapolates system metrics by a forecasted duration as specified in the configuration data input component
(10) based on the output obtained from the said regression analysis.
6. A system for load handling in cloud infrastructure as in Claim 3 wherein the breach conditions set up in the configuration data input component (10) by the user comprises at least a system metrics upper threshold (1 A) and a system metrics lower threshold (14B).
7. A system for load handling in cloud infrastructure as in Claim 1 or 2 wherein the said checking component (2C) performs condition checks by comparing predicted trend with system metrics thresholds set up in the configuration data input component (10) by the user.
8. A system for load handling in cloud infrastructure as in Claim 7 wherein the said checking component (2C) automatically launches at least one virtual machine (8) through the said provisioning server (4) when the predicted trend exhibits an increasing trend above the system metrics upper threshold (14 A).
9. A system for load handling in cloud infrastructure as in Claim 8 wherein the checking component (2C) registers additional new virtual machines (8) with the load manager (12) to handle application workload distribution.
10. A system for load handling in cloud infrastructure as in Claim 7 wherein the checking component (2C) automatically terminates at least one existing idle virtual machines (8) through the said provisioning server (4) when the predicted trend exhibits a decreasing trend below the system metrics lower threshold (14B).
11. A system for load handling in cloud infrastructure as in Claim 10 wherein the checking component (2C) deregisters existing idle virtual machines (8) from the load manager (12) to stop further workload distribution to the existing idle virtual machine (8).
12. A method of load handling in cloud infrastructure comprising steps of, gathering and storing system metrics historical information by monitoring agent (6); retrieving system metrics historical information from the said monitoring agent (6) by an analysing component (2 A); performing regression analysis on the system metrics historical information by the said analysing component (2A); using the output obtained from the regression analysis to extrapolate system metrics by a forecasted duration as specified in the configuration data in a predicting component (2B); performing condition checks in the checking component (2C) based on the output obtained from the said predicting component (2B) by comparing the predicted trend with system metrics thresholds set up in the configuration data input component (10); launching at least an additional new virtual machine (8) through a provisioning server (4) when the predicted trend exhibits an increasing trend above the system metrics upper threshold (14A) and registering the said additional new virtual machine (8); repeating the above step until the system metrics drops below the system metrics upper threshold (14 A); terminating at least an existing idle virtual machine (8) through the said provisioning server (4) when the predicted trend exhibits a decreasing trend below the system metrics lower threshold (14B) and deregistering the said existing idle virtual machine (8); repeating the above step until the system metrics rises above the system metrics lower threshold (14B).
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| MYPI2010006113 | 2010-12-22 | ||
| MYPI2010006113A MY178111A (en) | 2010-12-22 | 2010-12-22 | Intelligent load handling in cloid infrastructure using trend analysis |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2012087104A1 true WO2012087104A1 (en) | 2012-06-28 |
Family
ID=46314185
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/MY2011/000089 WO2012087104A1 (en) | 2010-12-22 | 2011-06-13 | Intelligent load handling in cloud infrastructure using trend analysis |
Country Status (2)
| Country | Link |
|---|---|
| MY (1) | MY178111A (en) |
| WO (1) | WO2012087104A1 (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB2494495A (en) * | 2011-09-07 | 2013-03-13 | Ibm | Determining virtual machine image pattern distributions in a networked computing environment |
| US20150294216A1 (en) * | 2014-04-10 | 2015-10-15 | International Business Machines Corporation | Cognitive distributed network |
| US9250944B2 (en) | 2011-08-30 | 2016-02-02 | International Business Machines Corporation | Selection of virtual machines from pools of pre-provisioned virtual machines in a networked computing environment |
| CN103810020B (en) * | 2014-02-14 | 2017-08-29 | 华为技术有限公司 | Virtual machine elastic telescopic method and device |
| US10353738B2 (en) | 2012-03-21 | 2019-07-16 | International Business Machines Corporation | Resource allocation based on social networking trends in a networked computing environment |
| US11579936B2 (en) | 2016-01-18 | 2023-02-14 | Huawei Technologies Co., Ltd. | System and method for cloud workload provisioning |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090300210A1 (en) * | 2008-05-28 | 2009-12-03 | James Michael Ferris | Methods and systems for load balancing in cloud-based networks |
| US20100223378A1 (en) * | 2009-02-27 | 2010-09-02 | Yottaa Inc | System and method for computer cloud management |
-
2010
- 2010-12-22 MY MYPI2010006113A patent/MY178111A/en unknown
-
2011
- 2011-06-13 WO PCT/MY2011/000089 patent/WO2012087104A1/en active Application Filing
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090300210A1 (en) * | 2008-05-28 | 2009-12-03 | James Michael Ferris | Methods and systems for load balancing in cloud-based networks |
| US20100223378A1 (en) * | 2009-02-27 | 2010-09-02 | Yottaa Inc | System and method for computer cloud management |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9250944B2 (en) | 2011-08-30 | 2016-02-02 | International Business Machines Corporation | Selection of virtual machines from pools of pre-provisioned virtual machines in a networked computing environment |
| US9465644B2 (en) | 2011-08-30 | 2016-10-11 | International Business Machines Corporation | Selection of virtual machines from pools of pre-provisioned virtual machines in a networked computing environment |
| GB2494495A (en) * | 2011-09-07 | 2013-03-13 | Ibm | Determining virtual machine image pattern distributions in a networked computing environment |
| US9038063B2 (en) | 2011-09-07 | 2015-05-19 | International Business Machines Corporation | Determining virtual machine image pattern distributions in a networked computing environment |
| US9317321B2 (en) | 2011-09-07 | 2016-04-19 | International Business Machines Corporation | Determining virtual machine image pattern distributions in a networked computing environment |
| GB2494495B (en) * | 2011-09-07 | 2016-10-05 | Ibm | Determining virtual machine image pattern distributions in a networked computing environment |
| US10353738B2 (en) | 2012-03-21 | 2019-07-16 | International Business Machines Corporation | Resource allocation based on social networking trends in a networked computing environment |
| CN103810020B (en) * | 2014-02-14 | 2017-08-29 | 华为技术有限公司 | Virtual machine elastic telescopic method and device |
| US20150294216A1 (en) * | 2014-04-10 | 2015-10-15 | International Business Machines Corporation | Cognitive distributed network |
| US10713574B2 (en) * | 2014-04-10 | 2020-07-14 | International Business Machines Corporation | Cognitive distributed network |
| US11579936B2 (en) | 2016-01-18 | 2023-02-14 | Huawei Technologies Co., Ltd. | System and method for cloud workload provisioning |
Also Published As
| Publication number | Publication date |
|---|---|
| MY178111A (en) | 2020-10-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2012087104A1 (en) | Intelligent load handling in cloud infrastructure using trend analysis | |
| US7568028B2 (en) | Bottleneck detection system, measurement object server, bottleneck detection method and program | |
| WO2019237593A1 (en) | Task processing method and apparatus, and computer device and storage medium | |
| WO2013107141A1 (en) | Policy scheduling system and method for cloud computing | |
| WO2005017736A1 (en) | System and program for detecting bottle neck in disc array device | |
| US9553810B2 (en) | Dynamic reconfiguration of network devices for outage prediction | |
| JP2006520027A5 (en) | ||
| EP2199915B1 (en) | Monitoring memory consumption | |
| JP2018532172A (en) | Method and system for resource scheduling | |
| US10305974B2 (en) | Ranking system | |
| US20200159559A1 (en) | Integrated configuration engine for interference mitigation in cloud computing | |
| US20140258382A1 (en) | Application congestion control | |
| JP6490806B2 (en) | Configuration method, apparatus, system and computer readable medium for determining a new configuration of computing resources | |
| US20140351414A1 (en) | Systems And Methods For Providing Prediction-Based Dynamic Monitoring | |
| EP3576359B1 (en) | Method and device for volume expansion of virtual network function | |
| US20190129813A1 (en) | System and a method for providing on-demand resiliency services | |
| US20170302588A1 (en) | Method and apparatus for provisioning of resources to support applications and their varying demands | |
| KR101608859B1 (en) | Cloud based intelligent power control system | |
| CN103944940A (en) | Equipment and method for dynamically configuring proxy server | |
| CN105791148B (en) | System and method for automatically balancing server load | |
| CN103095786B (en) | Online service request recognition methods, system, server and line server cluster | |
| KR20230055575A (en) | Universal large-scale multi-cloud environment monitoring system and method for private and public networks | |
| US20190347139A1 (en) | Technique for Determining a Load of an Application | |
| US9270742B2 (en) | Computer system and control method of computer system | |
| US8370839B2 (en) | Monitoring message queues in message queuing information systems and initiating batch jobs to perform functions on the message queues |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 11851778 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 11851778 Country of ref document: EP Kind code of ref document: A1 |